Subbasin Area/Base Flow Characteristics and Loss Rate Characteristics

HECHMS and Hydrologic Modeling

ENVI 512

February 2003

I.Hydrologic and Hydraulic Models

Representation of flooding is accomplished using two types of models: hydrologic and hydraulic. Hydrologic modeling simulates the hydrologic response (flow) of a basin to a given input of rainfall. HEC1 and HECHMS are two types of hydrologic models. Hydraulic modeling simulates the hydraulic response (water surface profiles) of a stream to a given input of flows. HEC2 and HECRAS are both hydraulic models. Both hydrologic and hydraulic models are needed for an in-depth flood analysis of any watershed system. While this exercise will only deal with using HECHMS for hydrologic modeling, a brief description of the four previously mentioned models are given below:

A. HEC1

The HEC-1 Flood Hydrograph package calculates discharge hydrographs for a given rainfall event. According to the U.S. Army Corps of Engineers, “The HEC1 model is designed to simulate the surface runoff response of a river basin to precipitation by representing the basin as an interconnected system of hydrologic and hydraulic components. …Representation of a component requires a set of parameters which specify the particular characteristics of the component and mathematical relations which describe the physical processes. The result of the modeling process is the computation of streamflow hydrographs at desired locations in the river basin.” Thus for computer modeling purposes, HEC1 divides the watershed into subwatersheds and reaches. Each subwatershed and reach uses averaged values over the area or stream length for the mathematical coefficients for the hydrologic and hydraulic computations. For this reason, HEC1 is referred to as a “lumped-parameter” model.

B.HECHMS

The U.S. Army Corps of Engineers Hydrologic Modeling System (HECHMS) is the Windows-based hydrologic model that supersedes HEC1 and contains many improvements over its predecessor. The most notable difference is an easy-to-use graphical user interface (GUI) shown in

Figure 1 which allows for the easy manipulation of hydrologic elements such as basin and river reaches and the easy input of basin characteristics. The GUI also allows for the easy viewing of results at any point in the model schematic, as shown in Figure 2. In the future, the model should have the capability to model gridded rainfall, such as NEXRAD-estimated rainfall.

Figure 1: HEC-HMS User Interface

Figure 2: Graphical Output from HEC-HMS

Another difference between HECHMS and HEC1 is the organization of the components that make up each hydrologic modeling run. In HECHMS, a project is created which contains separate “models”: the Basin Model, the Precipitation Model, and the Control Model. The user may specify different data sets for each model and then the hydrologic simulation is completed by using of data set for the Basin Model, the Precipitation Model, and the Control Model. The Basin Model contains the basin and routing parameters of the model, as well as connectivity data for the basin. The Precipitation Model contains the rainfall data, either historical or hypothetical, for the model. The Control Model contains all the timing information for the model, including model time steps and start and stops date and times of the simulation. This allows for easier organization of modeling data than in HEC1, which required a separate data set describing all aspects of the modeling run for each independent modeling run.

C.HEC2

The HEC2 Water Surface Profile program computes one-dimensional water surface profiles for steady flow in a stream with a given channel geometry. The program uses a backwater calculation to determine water surface levels starting from the given starting water surface elevation at the outlet at each profile based on peak flow, roughness (Manning’s) coefficients, cross-sectional geometry, bridges, culverts, and stream length. Like HEC1, the program is a DOS-based program.

D.HECRAS

HECRAS is the next-generation hydraulic model which supercedes HEC2. Like HECHMS, HECRAS features an easy to use GUI, and the ability to take advantage of HECDSS. Future versions of HEC-RAS will be able to model unsteady flow.

E.Other Modeling Tools

There are a number of tools that assist in the hydrologic and hydraulic modeling process. HEC-DSS and Visual DSS Basics 97 are presented below.

1.HECDSS

A major difference between HECHMS and HEC1 is the use of the Data Storage System, or HECDSS, to manage time-series and tabular data. The system was the result of a need in hydrologic engineering to relate similar types of data. Previously, data from one format would need to be entered into another format by hand by each user. Each program would then use separate functions to analyze and graph the data. The HECDSS software is the result of an effort to make hydrologic data management more efficient and allow for the HEC family of programs to use the same database.

What this entails is that time-series and tabular data are not stored in the HECHMS dataset; rather, the data are stored in a separate HECDSS data file, which is accessed by the HECHMS model. The database consists of six parts: the A Part, B Part, C Part, D Part, E Part, and F Part. The data are stored under a unique pathname, which includes all of the parts: /A Part/B Part/C Part/D Part/ E Part/F Part. Using these parts, it is easy for the user and the model to query and manage the data, especially between models.

Table 1: HEC-DSS Part Names

Part / Description
A Part / River basin or project name
B Part / Location of gage identifier
C Part / Data type (e.g. flow, rainfall, etc.)
D Part / Starting date
E Part / Time interval of data
F part / User defined descriptor of data

Long-term data series (years and greater) can be stored in HECDSS and multiple model runs can be made in different times within the data series. The data can also be accessed by other HEC models, such as HECFDA, which analyzes the cost-benefit of flood control and floodplain management alternatives.

2.Visual DSS Basics 97

While the HECDSS software package is an improved means of managing hydrologic data, there is not an easy method to transfer the data from a spreadsheet to a HECDSS file. A Microsoft Excel Add-In produced by Saracino-Kirby, Inc. titled Visual DSS Basics 97 allows the user to easily store regular time-series data from Excel into a HECDSS data-set. Visual DSS Basics 97 also allows for retrieval of data from the HECDSS database, which allows for the easy graphing of results in Excel with minimal user formatting.

II.Hydrologic Modeling and Parameter Estimation

A.Unit Hydrograph Methods (Runoff Transformations)

There are numerous methods of modeling runoff transformations for each subwatershed. We will present two of the more common methods, the Clark (TC+R) Unit Hydrograph and the Snyder Unit Hydrograph. Further discussion on both methods is presented in Hoggan (1997) and Bedient and Huber (1992).

1.Clark (TC+R) Unit Hydrograph

Need:

Time of Concentration (Tc): Hours
Storage Coefficient (R): Hours

Both the Brays Bayou and White Oak Bayou models use the Clark TC+R method for unit hydrograph computations. In theory, the Clark method uses a conceptual model consisting of a linear channel, which flows into a linear reservoir. Tc is the time of concentration, which can be measured in a gages basin as the time from the end of a burst of rainfall to its inflection point on the receding limb, normally measured in hours. Tc should also be the travel time from the most remote location in the watershed to the outlet. R, the storage coefficient, usually expressed in hours, can be estimated by taking the flow at the inflection point of the receding limb of the hydrograph and dividing it by the slope of the recession at the same flow. Figure 3 shows how to calculate the coefficients for Tc and R, based on an observed hydrograph.

Bedient and Huber (1992) present the equations for the determination of Tc and R:

, where

C=4295[% development]-0.678[% conveyance]-0.967, if % development > 18;

C=7.25 if % development is ≤ 18.

Figure 3: Calculation of Tc and R from an Observed Hydrograph
(adapted from Hoggan 1997)

, where C’ is taken from:

S0 (ft/mi) / % development / C’
> 40 / 0 / 5.12
20 < S0 ≤ 40 / 0 / 3.79
≤ 20 / 0 / 2.46
> 40 / 100 / 1.95
≤ 20 / 100 / 0.94

, where

L / = / length of channel (outflow to basin boundary) (mi)
Lca / = / length along channel to centroid of area (mi)
S / = / channel slope (ft/mi)
S0 / = / representative overland slope (ft/mi)
% development / = / percent of land that is developed (%)
% conveyance / = / ratio of flow to overland flow (%)
Tc / = / time of concentration (hr)
R / = / storage constant (hr)

Of these coefficients, determination of Lca is often the most confusing. To determine this coefficient, find the centroid of the subwatershed and draw the shortest line to the stream (say, point A). The distance from point A to the outlet along the length of the stream is Lca.

2.Synder Unit Hydrograph

Need:

Snyder’s Standard Lag (Tp): Hours
Snyder’s Storage Coefficient (Cp)

The Snyder method does not define a complete unit hydrograph, so the hydrologic model (HEC1 or HECHMS) completes the hydrograph using a trial and error procedure. With the given input parameters, Tpand Cp, the program uses hydrologic model to determine the optimal Clark parameters based on the Snyder coefficients. Bedient and Huber (1992) present equations to determine Tp and Cp:

, where

Tp / = / Snyder’s standard lag (hr)
L / = / length of channel (outflow to basin boundary) (mi)
Lc / = / length along channel to centroid of area (mi)
Ct / = / coefficient usually ranging from 1.8 to 2.2 (Ct has been found to vary from 0.4 in mountainous regions to 8.0 along the Gulf of Mexico)

Cp is the storage coefficient that normally ranges from 0.4 to 0.8, where larger values of Cp are associated with smaller values of Ct. Cp can also be estimated in the peak flow from the unit hydrograph is known:

, where

Qp / = / peak discharge of unit hydrograph (cfs)
A / = / drainage area (mi2)

B.Stream Routing

There are numerous methods of modeling stream routing for each stream segment in a hydrologic model. We will present two of the more common methods, Modified Puls (Storage-Outflow) Routing and Muskingum Routing. Further discussion on both methods is presented in Hoggan (1997) and Bedient and Huber (1992).

1.Modified Puls (Storage-Outflow) Routing

Need:

Storage: ac-ft vs. Outlflow: ft3/s
Time Steps (or subreaches)

Stream routing for the majority of the reaches in the Brays Bayou and White Oak Bayou watersheds is accomplished using the Modified Puls Method, or storage-outflow routing. The basis behind the Modified Puls Method is that the outflow in the channel is a unique function of storage in the channel. Calculation of the storage outflow relationship often involves a simple hydraulic computation of each river reach using a program such as HECRAS or HEC2. A table of corresponding storage and outflows is then entered into the model at each river reach.

Figure 4 shows how the relationship of storage and outflow is determined. The flow, Q, is a function of the water surface elevation, WS. The total storage, S, is then the simple geometric calculation of the volume of water between river segments A and B, and below the WS. Therefore, flow is a function of storage and these relationships are entered in the hydrologic model.

Figure 4: Modified Puls Stream Routing (adapted from Hoggan 1997)

Major complexities are determining the accurate storage in each reach. Often, idealized geometries for the model need to be assumed and simple open channel flow or more complex hydraulic analysis using HEC2 or HECRAS can be used to determine the associated flow.

The number of time steps is the time it takes a drop of water to travel the entire length of the routing reach divided by the computation time of the hydrologic model. To estimate the time it take a drop of water to travel the length of the reach, a hydraulic model should be used. As a rule of thumb, water in a stream can travel 2 mi/hr, although in channelized streams, the rate can increase to 10 mi/hr, or even greater, depending on overland slope and channel roughness.

2.Muskingum Routing

Need:

Muskingum K (Travel Time): hr
Muskingum X (Storage Routing)
Time Steps (or subreaches)

Muskingum K is the travel time for the reach, and is determined by dividing the mean velocity by the reach length. Velocity can be determined from a hydraulic model, such as HEC2 or HECRAS, or performing a simple open-channel flow calculation using Manning’s equation. Channel velocities can also be assumed, using the rule-of-thumb presented in the previous section.

Muskingum X is the only means represent storage for the routing step using this routing procedure. Muskingum X ranges from 0 to 0.5, where 0.5 is used for smooth uniform channels with a pure translation of the flood wave. A value of 0.2 is generally used for natural streams and a value of 0.45 is used for most improved urban channels.

Figure 5 shows the effect of Muskingum K and X coefficients on the routed hydrographs.

Figure 5: Muskingum Routing

C.Diversions

Diversions for hydrologic models use a simple table relating river flow to diverted flow. These relationships can be determined using geometric calculations and hydraulic models.

D.Loss Rates

The loss rate used in Harris County and many other communities is the simple Initial-Constant Loss Method. Under this method, an initial amount of rainfall is “lost,” or infiltrates (or evaporates) and a constant rate of rainfall is lost per hour. For Harris County, these rates range from 0.5 to 1.0 inches of initial loss, and 0.05 to 0.15 inches per hour of constant loss, depending of degree of urbanization and soil type (Table 2)

Table 2: HCFCD Recommended Losses

Sandy Soils / Clay Soils
Losses / Rural / Urban / Rural / Urban
Initial (in) / 1.00 / 0.75 / 1.00 / 0.50
Constant (in/hr) / 0.15 / 0.10 / 0.10 / 0.05

E.Baseflow

The baseflow method used in HEC1 is an exponential decay function of a defined starting baseflow according to the following equation:

where Qo is the starting baseflow, x is the empirical ratio between the recession baseflow and the recession baseflow one hour later (always greater than 1.0), and nt is the number of time steps. In simple hydrologic models over short time periods, baseflow can be neglected.

III.Example Problem

This example problem, will hopefully show the application of both unit hydrograph transformation methods and both routing methods. This problem is adapted from the U.S. Army Corps of Engineering HECHMS User Manual (Chapter 9). The user manual is available free of charge, as is the model, from their web-site This exercise will address the Castro Valley watershed, a small basin located in northern California, and its response to a rainfall event in 1973.

A.Background information

The Castro Valley example problem uses a small watershed, consisting of four subbasins and two stream reaches which will be modeled as routing steps (Figure 6). Following is a description of all the data necessary to setup HMS for this watershed, and a list of commands needed to do the setup.

B.HECHMS Model Set-Up (Getting Started)

This exercise will recreate a HECHMS model from scratch for the Castro Valley watershed in northern California.

Open HECHMS program by double-clicking on the HMS icon.
To create a new project, select New in the pop-up window. Name the project Castro Example and type Castro Valley urban study for the description. All information will be, by default, stored in the folder in the C:\hmsproj\Castro Example folder. Click OK.
To set the defaults for the new project and the desired unit system, select Project Attributes from the File menu. Select English units, Initial/Constant loss rate, Clark UH, and Muskingum routing, and then click OK to save the defaults.

C.Gage Data

There are three ways to enter gage (precipitation and flow data) into the HECHMS model. The first way is to directly enter the data. The second way is to import the data from an existing DSS file. The last way is to import the data from Excel using Visual DSS Basics 97.

To enter gage data by the user, select Precipitation gages under the Data menu. The first time you use this, a new gage will automatically be created. For additional gages, you will choose Add Gage from the Edit menu. Enter Fire Dept. under Gage ID, and make sure that Incremental Precipitation and Inches are selected in the Data Type and Units windows, respectively. Select Manual Entry and click OK. Enter 16JAN73 and 0300 for starting date and time (January 16, 1973 at 3:00 AM) and enter 16JAN73 and 1000 for ending date and time. Select 10 Minute time interval and click OK. Enter the rainfall at each time step from Table 3 and click OK and Close.

Table 3: Fire Dept. Rainfall on 16JAN73
(Time is end of rainfall total)

3:10 / 0.00 / 4:20 / 0.03 / 5:30 / 0.03 / 6:40 / 0.07 / 7:50 / 0.03 / 9:00 / 0.00
3:20 / 0.00 / 4:30 / 0.02 / 5:40 / 0.09 / 6:50 / 0.07 / 8:00 / 0.01 / 9:10 / 0.01
3:30 / 0.01 / 4:40 / 0.05 / 5:50 / 0.08 / 7:00 / 0.02 / 8:10 / 0.03 / 9:20 / 0.06
3:40 / 0.01 / 4:50 / 0.05 / 6:00 / 0.03 / 7:10 / 0.04 / 8:20 / 0.02 / 9:30 / 0.02
3:50 / 0.08 / 5:00 / 0.02 / 6:10 / 0.04 / 7:20 / 0.03 / 8:30 / 0.01 / 9:40 / 0.04
4:00 / 0.03 / 5:10 / 0.05 / 6:20 / 0.03 / 7:30 / 0.02 / 8:40 / 0.03 / 9:50 / 0.01
4:10 / 0.05 / 5:20 / 0.04 / 6:30 / 0.07 / 7:40 / 0.03 / 8:50 / 0.01 / 10:00 / 0.00

OPTIONAL: To see how Visual DSS Basics 97 works, close down HECHMS. We have to do this because to programs cannot access the same DSS file at the same time. Open up Excel, and click I Agree for the Visual DSS Basics 97 dialog box. In Excel, click Open… under the Visual DSS Basics 97 File menu (which should be below the File menu for Excel or in a separate dialog box. Explore to the hmsproj file, open the Castro2 folder, double-click on Castro2.dss file, and click OK. This now allows access to the DSS dataset. In the Visual DSS Basics 97 menu, click Retrieve and then Regular Time Series and By Selection…. Click on Full and click OK. If your DSS dataset has become too large, you may narrow the search using different qualifiers for the different parts using the Selective search. Under Time Window, enter the starting date time for the series (January 16, 1973 at 3000) and the ending data and time (January 16, 1973 at 1000). Scroll down until you find the DSS file for the Fire Dept. gage, which should be at something like: /CASTRO VALLEY/FIRE DEPT./PRECIP-INC/16JAN73/10MIN/OBS/. Highlight the DSS pathname and click OK. You can make any changes to the data that you wish inside of Excel, or import new data and change the name of the gage in Part B. To store the new data, simply highlight from the lower right hand corner of the data to Part A: and click on Store and then Regular Time Series. At the dialog box, click OK and the new data should be stored in the DSS file. Close the DSS file by clicking File on the Visual DSS Basics 97 menu and Close. You can save the worksheet if you, for easy import later into HECHMS. DSS files imported into the DSS database using Visual DSS Basics 97 needed to have a gage identified with the DSS file. To do this, in the main menu of HECHMS, click on Edit – Gage Data – Precipitation. In the pop-up menu, click on Add Gage and give the gage a name in the Gage ID field. Click on External DSS File and click OK. In the next pop-up menu, click on Generate Catalog (again, you can limit your query by using the different parts). Find the gage and highlight its pathname and click OK and Close.
The final way to input gage data into the HECHMS model is to retrieve data from an existing DSS dataset. We will do this for the observed flow at the outlet of the watershed. Select Data – Discharge Gage. The first time you use this, a new gage will automatically be created, after that, you will have to select File – Add Gage for a new gage. Name the gage Outlet and select External DSS Record, then click OK. The DSS file is actually in the existing Castro file, so click on File Browser in the upper right hand corner of the window to select the right DSS file. Explore to the castro folder in the hmsproj folder and open castro.dss. Click on Generate Catalog and highlight the pathname for /CASTRO VALLEY/OUTLET/FLOW/16JAN73/ 10MIN/OBS/, click OK and then Close. You have now imported an existing DSS file from another program into a new dataset.

D.Basin Model